TW571108B - Double-pulsed optical interferometer for waveform probing of integrated circuits - Google Patents

Abstract

Optical interferometery is used to probe an integrated circuit device under test (DUT). During each cycle of a repetitive electrical test pattern applied to the DUT a reference pulse is provided at a fixed time relative to the test pattern, and a probe pulse is provided at a time scanned through the test pattern in the manner of equivalent time sampling. The probe and reference light pulses are each split to provide at least a second probe pulse and a second reference pulse. One probe pulse and one reference pulse interact with the DUT at the same physical location, but at displaced times with respect to each other. The second probe pulse and the second reference pulse travel an optical delay path with length controlled to compensate for motions of the DUT. The probe pulses are recombined and detected to provide a probe interference signal. The reference pulses are recombined and detected to provide a reference interference signal. For each cycle, electrical activity in the DUT is detected by preparing a ratio of the probe interference signal and the reference interference signal.

Description

571108 五、發明說明α) 相關申請案交叉參考 本申清案主張1 9 9 9年9月24日提出之美國專利臨時申 請案第6 0 / 1 5 5 8 4 4號的益處，該臨時申請案在此納入來 本發明關於積體電路之光學測試，特 活動之光干涉儀探測。 销股电塔安571108 V. Description of the invention α) Cross-reference to related applications This application claims the benefits of US Patent Provisional Application No. 6 0/1 5 5 8 4 4 filed on September 24, 1999. This provisional application This case incorporates the optical testing and integrated optical interferometer detection of the integrated circuit of the present invention. Taian

積體電路之電活動能藉由以光學方式探測因改變 體二極體接合處之偏壓和載體密度引起之折射率和吸V ，變化而監測。舉例來說，半導體材料之折射率 J =化能藉由自二極體接合處反射之光束密度的變化得； 入本：： = p/nlccla)之美國專利第5872360號"内 一實施例中，一雷射束以接近一 =之電场。在 隙的波長作用。該Φ射击办屯Γ 脰 ° 之能帶 P-N (二極體）接人田/ (彳，牙匕+導體基底之背側聚焦在〜 之汲極區）上^Λ由 金屬氧化物半導體電晶體 氧化物界面及金屬;束ΪΓ接ί處’在接合處背後之 而。拉入老 射然後折返牙過接合處並離開石夕矣 ° a处内電極之調變使此反射光束因電吸收和電拼2 之總應造成振幅調變。 电折射 太八^:丄^咖^等人之美國專利第5 9 0 5 5 7 ^^ 受;# t $跤揭不積體電路之雙雷射束探測。在施加於 測光束對該受測：=試圖信號的每一循環ΐ利用-探 、 牛上的波形取樣。亦使用一參考雷射束 第6頁 571108 五、發明說明（2) 在與該探測光束相同之取樣物理位置對該受測元件取樣 每一參考測量皆相對於測試圖為一固定時間，而探剛測量 係以正交於成等時取樣之方式知過測試圖以重新建構、皮 形。對每一測試循環取得探測測量和參考測量之比例以 低棟測測量因雜訊造成的波動。 + 因一受測元件之電活動造成的折射率和吸收係 亦能藉由一透射通過受測元件或反射 ’、▲艾化 調變會比探測折射率與受測元件電S 關係更為靈敏。亨利奇（Heinrich ) 、，卞电活動 ⑽0 9 2號（納入本發明參考）揭示人之吳/專利第 體7G件造成相位調變之干涉測量方 一又主動半導 光束且皆聚焦在受測元件上。—/ 。一光束分流為兩道 上’該光束在此由—調變折射 1 ^ :焦在-活性區 光束聚焦在一非活性區以提供泉相位調變，而另一道 涉。干涉光束之強度調變歸因於因、I 5射的光束重組並干 測光束相位調變。 、文測區電活動造成之探 '一般而古 ，T > , > 0 干涉測量方法對；^ 感。在予利奇等人之方法 了振動和溫度變化極端敏 上來說對與光束軸線一致之為兩干涉光束之相對相位名義 為不敏感。然而在〜古 ^及凡件振動及溫度誘發運動 光束的適當點。同：仍：：體電路上报難找到放參考 感。 I崎越雷射束轴線之運動為敏 因此需要一錄 對受測元件之振動形探測之干涉測量方法，其 _— 酿度誘發運動不敏感，且不用在受測The electrical activity of the integrated circuit can be monitored by optically detecting changes in the refractive index and the absorption V caused by changing the bias voltage and carrier density at the junction of the body diode. For example, the refractive index of a semiconductor material, J = chemical energy, is obtained by changing the density of the light beam reflected from the junction of the diode; in this example: = p / nlccla) US Patent No. 5872360 " an embodiment In a laser beam, an electric field close to one =. Wavelength effect in the gap. The Φ firing office Γ 脰 ° band PN (diode) is connected to the field / (彳, the back of the tooth + conductor substrate is focused on the ~ drain region) ^ Λ by a metal oxide semiconductor transistor The oxide interface and the metal; the beam ΪΓ is connected behind the joint. Pull in the old beam and then return to the tooth and pass through the joint and leave Shi Xi 矣 ° The modulation of the internal electrode at a makes the reflected light beam amplitude modulation due to the electric absorption and electric spelling 2. Electrical refraction Taiba ^: 丄 ^ Ca ^ et al. US Patent No. 509 0 5 7 ^^; After the test beam is applied to the test: = each cycle of the attempted signal, the waveform on the probe is sampled. A reference laser beam is also used. Page 6 571108 V. Description of the invention (2) The test element is sampled at the same sampling physical position as the detection beam. Each reference measurement is a fixed time relative to the test pattern, and the probe The new measurement system has known the test chart in a way orthogonal to the isochronous sampling for reconstruction and skin shape. For each test cycle, the ratio of the probing measurement to the reference measurement is taken to measure the fluctuation caused by noise. + The refractive index and absorption system caused by the electrical activity of a device under test can also pass through the device under test or reflect through a transmission. The Ai modulation will be more sensitive than detecting the relationship between the refractive index and the electrical S of the device under test. . Heinrich, 卞 Electrical Activity No. 0 92 (incorporated in the present invention reference) reveals that the Wu / Patent No. 7G piece caused phase modulation of the interferometric measurement. Both active semiconducting beams are focused on the test. Component. — /. A beam splits into two paths. The beam is here refracted by modulation. 1 ^: Focused on the active region. The beam is focused in an inactive region to provide spring phase modulation, while the other is involved. The intensity modulation of the interference beam is attributable to the recombination of the beam emitted by I 5 and the measurement of the beam phase modulation. 3. Probe into the Cause of Electrical Activities in the Wenxian District 'General and Ancient, T >, > 0 Interference measurement method pairs; In the method of Yu Liqi et al., Which is extremely sensitive to vibration and temperature changes, it is nominally insensitive to the relative phase of the two interference beams, which is consistent with the beam axis. However, in the ancient times and where the vibration and temperature-induced motion of the pieces are appropriate. Same: Still :: It is difficult to find the reference signal on the body circuit. The motion of the Izaki laser beam axis is sensitive, so an interferometry method for detecting the vibration shape of the component under test is needed. Its _— the degree-induced motion is not sensitive, and it does not need to be measured.

$ 7頁 571108 五、發明說明（3) 元件上之相關活性 一種在重複電 元件之電活動的方 區附近找 測試圖信 中於一選定 之每一重複 移的時間提 一參考光脈 二參考光脈 至半導體元 在該第 相互作用之 使二者之空 考光脈衝組 衝受偵測提 出一適當參考點。 於半導體元件時偵測該 該電測試圖之每一重複 脈衝，且在該電測試圖 一探測光之選定時間遷 該第一探測光脈衝和第 分流提供至少一第二探測光脈衝和一第 第一探測光脈衝和第一參考光脈衝導向 號施加 包括在 時間提供一第一探測光 中於一相對於提供該第 供一第一參考光脈衝。 法，其中 衝各自 衝。該 件之一 一探測 區域上。 光脈衝和 測提供一參考干涉 異。探測干涉信號 複數個選定時間決 後，該第一探測 間和時間重疊， 合使二者之空間 供一探測干涉信 信號。所 與參考干 定 第一參考光脈衝與半導體元件 光脈衝與第二探測光脈衝組合 且該第一參考光脈衝與第二參 和時間重疊。重疊的探測光脈 號，且重疊的參考光脈衝受偵 選時間隨電測試圖之重複性而 涉信號之比例在電測試圖内之 光脈衝源 器 種相關裝置包括 器， ，及 分$ 7 pages 571108 V. Description of the invention (3) Relevant activity on the component A type of test chart is found near the square area of the electrical activity of the repeated electrical component. At a selected time for each repeated movement, a reference is made to the light pulse. The light pulse to the semiconductor element at the first interaction causes the empty light pulse group of the two to be detected and propose an appropriate reference point. Detecting each repetitive pulse of the electrical test pattern at the time of the semiconductor element, and migrating the first detection light pulse and the first shunt at a selected time of a detection light of the electrical test pattern to provide at least a second detection light pulse and a first The application of the first detection light pulse and the first reference light pulse guide sign includes providing a first detection light at a time relative to providing the first reference light pulse with the first supply. Method, where the punches are punched separately. One of the pieces is on the detection area. Light pulses and measurements provide a reference interference difference. Detecting the interference signal After a plurality of selected time decisions, the first detection interval and time overlap, so that the space between the two provides a detection interference signal. With reference, the first reference light pulse and the semiconductor element light pulse are combined with the second detection light pulse, and the first reference light pulse and the second reference time overlap. Overlapping detection light pulses, and the overlapping reference light pulse selection time depends on the repeatability of the electrical test chart. The ratio of the signal involved in the electrical test chart is the optical pulse source device.

/；IL 器系統 號比例之處理器。 偵測 第 半 判 探測光脈衝源 第一參考 導體元 定探測 第一圖為依據一本發明實施例 第二圖為第一圖裝置之局部放 第三圖繪出解釋第一圖裝置作 件支撐器，一光束組合 干涉信號與參考干涉信 之裝置的簡圖。 大圖。 業之干涉信號。/; IL system processor. Detect the first half of the detection light pulse source. The first reference conductor element determines the detection. The first picture is based on an embodiment of the present invention. The second picture is a part of the device of the first picture. The third picture is used to explain the device support of the first picture. A simplified diagram of a device that combines a beam of interference signals with a reference interference signal. Big picture. Interference signal of industry.

第8頁 571108 五、發明說明（4) ___ 心利用光干涉測量技術探測積體電路。在〜 =測積體電路元件（DUT)之電測試圖信號'^重複施加於 %中，提供—探測光脈衝及一參考光脈衝且、_母一重複循 光學路徑以對該受測元件上之電波形取樣。=者行經相同 的持續時間比測試圖内相關最高頻信號之週^等探測脈衝 ^ 大τ覓測試圖波形測量。該等參考脈衝短，藉以提 二比探測脈衝長。參考脈衝係相對於測試圖持續時間通 曰供。探測脈衝在一系列測試圖循環中以正六固定時間 之方式掃過測試圖以重新建構受測元件波形=於等時取樣 部分以 探測脈 第二探 第一參 分流並偵測每一探測脈衝和每一參考脈衝 =出入射於受測元件上之脈衝能量的特徵。之 1丨ΐ二干涉儀内/分流提供至少-第-探測脈衝2 去,且芩考脈衝在一干涉儀内分流提供小— 脈衝和一第二參考脈衝。第―探測脈衝=:第 文測元件上之波形於相同物理位置取樣，二苓考脈衝 衝掃過測試圖之一變動延遲而錯開時間 ~错由探測脈 弟二參考脈衝在干涉儀之一延遲一彳采測脈衝和 經相同路徑。 、遲支路（delay arm )内行 ，第一探測脈衝在與受測元件相互 脈衝重組而在一偵測器提供一探 \'、弟二探測 -參考脈衝在與受測元件相互作用地，第 組而在一偵測器提供—參考灸/、弟—芩考脈衝重 使受測元件内電活動變動（苴〜三在探測脈衝因時間 相互作用時調變第一探測脈衝：:罙測脈衝與受測元件 尤予相位）而依時間掃過 第9頁 571108Page 8 571108 V. Description of the invention (4) ___ The heart uses the optical interferometry to detect the integrated circuit. The signal of the electrical test pattern signal of the test object circuit element (DUT) is repeatedly applied in%, and the detection light pulse and a reference light pulse are provided, and the mother and the mother repeatedly repeat the optical path to the device under test. Electrical waveform sampling. = Those who travel by the same duration than the detection pulse of the highest frequency signal in the test chart ^ are larger than the detection pulse ^ find the test chart waveform measurement. These reference pulses are short, so that they are longer than the detection pulses. The reference pulses are given relative to the duration of the test pattern. The detection pulse is scanned through the test pattern in a series of regular six fixed times in a series of test pattern cycles to reconstruct the waveform of the device under test. The isochronous sampling portion is used to detect the pulse. Each reference pulse = characteristic of the pulse energy incident on the device under test. One of the two interferometers in the shunt / shunt provides at least -the first detection pulse 2 to go, and the test pulse shunts in the interferometer to provide a small-pulse and a second reference pulse. The first-detection pulse =: The waveform on the second test element is sampled at the same physical location. The Erling test pulse sweeps through one of the test patterns and the delay is delayed and staggered. The pulses are taken at the same time and through the same path. In the delay arm, the first detection pulse is recombined with the device under test to provide a probe in a detector. The second detection-reference pulse interacts with the device under test. The test pulses provided by a detector—reference moxibustion, and younger brother—revisited the electrical activity within the tested element (苴 ~ 三, the first detection pulse is modulated when the detection pulse interacts with time :: 罙 测 脉Especially in phase with the component under test) and swept over time according to page 571108

受測元件橫跨光學軸線之溫度變異和運動同樣會影響 探測反射比和參考反射比。源自此等來源之雜訊藉由計算 出探測反射比和參考反射比而消去。探測反射比與參考反 射比之比例的剩餘調變代表受測元件内電活動之波形。The temperature variation and movement of the measured component across the optical axis will also affect the detection reflectance and reference reflectance. Noise from these sources is eliminated by calculating the detection reflectance and the reference reflectance. The remaining modulation of the ratio of the detection reflectance to the reference reflectance represents the waveform of the electrical activity within the device under test.

第H)頁 571108Page H) 571108

第一圖為一雙脈衝光干涉儀系統6 0 0之簡圖。位在習 知雷射平台6 04上之一鎖模雷射6〇2輸出雷射脈衝序列 6 0 6。在一實施例中，該鎖模雷射為Ughtwave Electronics Model· 13卜2 0 0鎖模歛：釔鋁石榴石雷射 (Nd: YAG laser )供應約35微微秒（ps )持續時間之 100MHz重複率、1 0 64毫微米波長脈衝。脈衝序列6S〇6通過 光調變器60 8，該光調變器為一重複施加於受測元件之測 試圖的每一循環自脈衝序列606選擇一探測脈衝6〇9 (如 610處所示）。光調變器608例如為一具備c〇n〇ptics 25〇 驅動件之Conptics 36 0-80 Electro-Optic Modulator， 然亦可使用其他光調變器如聲光調變器。在一實施例中， 光調變器6 0 8包括二個串列光調變器以更為完全抑制脈衝 序列6 0 6内不要的脈衝。 連續波雷射6 1 2亦位於雷射平台6 0 4上且對光調變器 6 1 6輸出一連續振幅雷射束6 1 4。在一實施例中，連續波雷 射612 為Coherent, Inc. CompassTM Model 1064-500 連續 波鈥：釔鋁石榴石雷射，該雷射亦以1 〇 6 4毫微米之波長工 作。光調變器616例如也是Conptics 3 6 0 -8 0 Electro-Optic Modulator，然亦可使用其他光調變器。 在一實施例中，光調變器6 1 6包括二個並列光調變器。光 調變器6 1 6調變光束6 1 4對一施加於受測元件之測試圖的每 一循環提供一持續時間短達1 0毫微秒（n s )之參考脈衝 6 1 8。在一實施例中，該參考脈衝之持續時間為1 5 0毫微 秒，且參考脈衝與探測脈衝以一約1 · 5微秒（s )至約1The first figure is a simplified diagram of a dual-pulse optical interferometer system 600. One of the mode-locked lasers 602 on the conventional laser platform 604 outputs a laser pulse sequence 606. In one embodiment, the mode-locked laser is a Ughtwave Electronics Model. 13 2 0 0 mode-locked: yttrium-aluminum-garnet laser (Nd: YAG laser) supplies a 100 MHz repetition with a duration of about 35 picoseconds (ps) Rate, 1064 nm wavelength pulse. The pulse sequence 6S〇6 passes through the optical modulator 60 8 which selects a detection pulse 609 from the pulse sequence 606 for each cycle of a test pattern repeatedly applied to the device under test (as shown at 610). ). The optical modulator 608 is, for example, a Contics 36 0-80 Electro-Optic Modulator with a cnoptics 25〇 driver, but other optical modulators such as acousto-optic modulators can also be used. In one embodiment, the optical modulator 608 includes two tandem optical modulators to more completely suppress unwanted pulses in the pulse sequence 606. The continuous wave laser 6 1 2 is also located on the laser platform 6 0 4 and outputs a continuous amplitude laser beam 6 1 4 to the light modulator 6 1 6. In one embodiment, the continuous wave laser 612 is a Coherent, Inc. CompassTM Model 1064-500 continuous wave ': yttrium aluminum garnet laser, which also operates at a wavelength of 106 nanometers. The optical modulator 616 is also, for example, a Contics 3 6 0-8 0 Electro-Optic Modulator, but other optical modulators can also be used. In one embodiment, the light modulator 6 1 6 includes two parallel light modulators. The light modulator 6 1 6 modulates the light beam 6 1 4 for each cycle of a test pattern applied to the device under test. A reference pulse 6 1 8 of a duration as short as 10 nanoseconds (n s) is provided. In one embodiment, the duration of the reference pulse is 150 nanoseconds, and the reference pulse and the detection pulse are within a range of about 1.5 microseconds (s) to about 1

第11頁 571108 五、發明說明 的ι日寸就日守間分開，該延時視受測元件電測試 毫秒（ms 圖而定 板測脈衝和參考阶彳私 CVI Laser Γη 亏脈衝通向一光束組合器620，例如— C〇rP〇rati- ^WSl-l〇64-10-l〇25-45P^r 束組合器620將探測脈衝和\\轉光學件如反射鏡622。光 偏振器623至光纖耦接哭脈衝沿共線路徑導向通過 參考脈衝耦合於維持偏振單r Τ光纖耦接器將探測脈衝和 分流比選擇為在= 纖62 6。光束組合器62。之 6 20弄成共線之德，失本和翏考脈衝之路徑由光束組合器 量。偏振器623將探測脈^ ^ ^約等於探測脈衝能 態。 衝和翏考脈衝設定於相同偏振狀 另種述擇，探測脈衝和 _ ^ ^ 如單-鎖模雷射）提供。在“考脈：係由相同光源（例 二光束分流器分流為探測脈：心：衝早-輸出脈衝由 施例令，探測脈衝和參考脈衝：=供-延遲。在另-實 ，立光脈衝。在其他實施例中使用、相同雷射輸出之 二相干光源如發光二極體作為提：：射以外之雷射及 衝。 ’、提ί、板測脈衝和參考脈 光纖6 2 6將探測脈衝和參考 ’該耦接器將該等脈衝耦接於柃=引至光纖耦接器 」偏振光束分流器632上。光纖626^田系統630且在此入射 參考脈衝空間濾波，藉此使二 &、先天性對探測脈衝和 二間模結構大致相符。 571108 五、發明說明（8) 每一探測脈衝釦会+ 分流器632至光/_考脈衝之一線性偏振部分通過偏振光束 儀_)。以；：：：=(其將該等脈衝轉入干i 其㈣來為-受測元件造像。一、“亀顯微鏡’ 圖。干：儀6 9 9 ΐ其相關控制電子系統之詳細 波片71 〇、孔徑&秀7^ ί路由反射鏡705、四分之一 Γ=:分之—波片63 9、物綱及受二元件心 ;二；：中受測元件640由—習知η臺⑷支擇在 只 中，干涉儀6 9 9和掃描系統63 0由一χ — y —ζ Α古 ΐη於一靜止受測元件64°運動。光束635内之探至制 ▲Γ广人考脈衝618 a射於非偏振'束分流器701 i將每」 入、&衝分流為二道脈衝。為了方一' _藉由光束分流議傳入受測元件支路内：部 測脈衝6〇9a，且每一探測脈衝轉入干涉儀699 L遲土路内之部分則稱為第二探測脈衝6 09b。相似地，每 一茶考脈衝618藉由光束分流器7〇1傳入受測元件支路 部分在以下稱為第一參考脈衝6183，且每一參考脈衝轉入 干涉，699延遲支路内之部分則稱為第二參考脈衝61扑。 第一探測脈衝6 0 9a和第一參考脈衝61仏通過四分之一 波片6 3 9，該片通常定向為將脈衝之偏振性從線性轉換為 圓形。圓偏振第一探測脈衝6 093和第一參考脈衝618&由物 鏡636聚焦在受測元件640活性區63 8内之相同點。每一脈 571108 五、發明說明（9) 衝之一部分自受測元件反射回來通過物鏡636 (其重新校 準光束）至四分之一波片6 3 9。四分之一波片6 3 9將脈衝之 偏振性從圓形偏振轉換為與原始線性偏振正交之線性偏 振。在通過四分之一波片6 3 9之後，每一反射的第一探測 脈衝和第一參考脈衝的一部分通過光束分流器7〇 1回到光 束路徑63 5入射於掃描系統6 3 0上。 弟一棟測脈衝6 0 9 b和第二參考脈衝6 1 8 b由光束分流器 701轉向通過快門703沿干涉儀延遲支路光束路徑7〇4行 進。快門7 0 3在系統用於干涉測量之時為開放。為了方便 起^ ’光束路徑704由反射鏡70 5折向。第二探測脈衝⑽讣 和第二參考脈衝6丨8b通過四分之一波片7丨〇和孔徑7丨2至透 鏡71 5 (其將該等脈衝聚焦於反射鏡72〇上）。安裝在壓電 致動器72 5上之反射鏡720的位置調整為將脈衝沿其入射路 徑反射回去。反射的脈衝往回通過透鏡715、孔徑Μ?及四 分之一波片710至反射鏡70 5 (其將脈衝轉向通過快 至光束分流器70"。四分之一波片71〇定向為使‘次通過 四分之一波片710之反射脈衝線性偏振正交於原始偏 與反射的第-探測脈衝和第一參考脈衝相#。孔徑' 整反射探測光束和參考光束在延遲支路内之 厂一 件支路内之反射探測光束和參考光束直徑相符”叉濟兀 每一反射的第二探測脈衝和第_ A 光束分流器m轉回沿光束路徑= m 脈衝和第-參考脈衝之部分回到 + 射的弟—探測 分流™亦作用為一光束組合器=糸一=Page 11 571108 V. The description of the invention is separated from the day by day. The delay depends on the electrical test of the device under test in milliseconds (ms chart and the plate test pulse and the reference order. CVI Laser Γη Defective pulse leads to a beam combination 620, for example-C〇rPrati- ^ WSl-l〇64-10-l〇25-45P ^ r Beam combiner 620 will detect the pulse and turn optics such as mirror 622. Light polarizers 623 to The optical fiber coupling pulses are guided along a collinear path, and the reference pulse is coupled to the maintaining polarization single r T fiber coupler. The detection pulse and shunt ratio are selected to be = fiber 62 6. beam combiner 62. of 6 20 to be collinear. The path of the loss, loss, and test pulse is measured by the beam combiner. The polarizer 623 sets the detection pulse ^ ^ ^ approximately equal to the energy state of the detection pulse. _ ^ ^ As single-mode-locked laser). In "Examination pulse: the same light source (example two beam splitter shunts as the detection pulse: heart: rush early-output pulse by the order, detection pulse and reference pulse: = supply-delay. In another-real, Liguang Pulses. In other embodiments, two coherent light sources with the same laser output, such as light emitting diodes, are used as lasers: lasers and pulses other than lasers. ', 、, plate test pulses and reference pulse fibers 6 2 6 will The detection pulse and the reference 'the coupler couples these pulses to the 柃 = lead to the fiber coupler' polarization beam splitter 632. The optical fiber 626 ^ field system 630 and the reference pulse spatial filtering are incident here, thereby Second, the congenital pair of detection pulses and the two-mode structure are roughly consistent. 571108 V. Description of the invention (8) Each detection pulse buckle + shunt 632 to one of the light / test pulses The linear polarization part passes through the polarizing beam meter _). Take ::: == (It transfers these pulses into the stem, and then it creates an image of the device under test. I. “亀 microscope” picture. Stem: instrument 6 9 9 ΐ detailed wave plate of its related control electronic system 71 〇 Aperture & Show 7 ^ ί route mirror 705, one quarter Γ =: one-half-wave plate 63 9, physical outline and the second element center; two ;: middle test element 640 by-the known η is the only option, the interferometer 6 9 9 and the scanning system 63 0 are moved by a χ — y ζ Α 古 ΐη at 64 ° on a stationary test element. The detection system in the beam 635 The test pulse 618 a is shot at the non-polarized 'beam splitter 701 i. Each pulse will be split into two pulses. For the sake of Fang Yi' _ is introduced into the branch of the device under test through the beam splitting method: internal test pulse 609a, and the portion where each detection pulse is transferred into the interferometer 699 L late soil road is called the second detection pulse 6 09b. Similarly, each tea test pulse 618 is transmitted to the receiver through the beam splitter 701. The part of the measuring element branch is hereinafter referred to as the first reference pulse 6183, and each reference pulse turns into interference, and the part in the 699 delay branch is called the second reference pulse 61 fl. The first detection pulse 6 0 9a and the first reference pulse 61 仏 pass through a quarter wave plate 6 3 9 which is generally oriented to convert the polarization of the pulse from linear to circular. The first detection pulse 6 093 is circularly polarized It is the same as the first reference pulse 618 & focused by the objective lens 636 in the active area 638 of the test element 640. Each pulse 571108 V. Description of the invention (9) A part of the impact is reflected back from the test element and passes through the objective lens 636 (its Recalibrate the beam) to the quarter wave plate 6 3 9. The quarter wave plate 6 3 9 converts the polarization of the pulse from circular polarization to linear polarization orthogonal to the original linear polarization. After a wave plate 6 3 9, a part of each reflected first detection pulse and first reference pulse returns to the beam path 63 5 through the beam splitter 70 1 and is incident on the scanning system 6 3 0. 6 0 9 b and the second reference pulse 6 1 8 b are diverted by the beam splitter 701 along the interferometer delay branch beam path 704 through the shutter 703. The shutter 703 is open when the system is used for interferometry. For convenience, the beam path 704 is 50% off by the mirror 70 The second detection pulse ⑽ 讣 and the second reference pulse 6 ˜ 8 b pass through the quarter wave plate 7 ο and the aperture 7 丨 2 to the lens 71 5 (which focuses the pulses on the mirror 72). The position of the mirror 720 mounted on the piezoelectric actuator 72 5 is adjusted to reflect the pulse back along its incident path. The reflected pulse passes back through the lens 715, the aperture M? And the quarter wave plate 710 to the mirror. 70 5 (It turns the pulses through to the beam splitter 70 ". The quarter wave plate 71 is oriented such that the linear polarization of the reflected pulses of the 'pass through quarter wave plate 710' is orthogonal to the first detection pulse and the first reference pulse phase # of the original polarization and reflection. Aperture 'The entire reflection detection beam and reference beam are in the delay branch. The reflected detection beam in one branch matches the diameter of the reference beam. "The second detection pulse and the _A beam splitter m for each reflection. Turn back along the beam path = part of the m-pulse and the -th reference pulse back to + the younger-probe shunt ™ also acts as a beam combiner = 糸 一 =

571108571108

分流器7 0 1之分流比選擇A , , t ^ ^ 射信號…反射的第上一來自文測元件640之典型反 射）能量對反射的第-夂：考脈衝（自受測元件640反 日L… 弟一翏考脈衝（自反射鏡72 0反射）能 ，匕例在離開干涉儀時約為！ ：3。在另-實施例中，八 :比4擇為使反射麥考脈衝能量的比例約為i : i。： ::例中，反射參考脈衝能量之比例藉由在延遲支路路徑 内***—功率衰減器（例如—中性濾、光片）調整。The split ratio of the shunt 7 0 1 selects A,, t ^ ^ The transmission signal ... The first reflected typical reflection from the text measurement element 640) The first-夂 of the energy to reflection: test pulse (self-test element 640 anti-day L … I can test the energy of the pulse (reflection from the mirror 72 0), when the dagger is away from the interferometer !: 3. In another embodiment, the ratio of 8: 4 is selected as the ratio of the energy of the reflected McCaw pulse. Approximately i: i .: ::: In the example, the ratio of the reflected reference pulse energy is adjusted by inserting a power attenuator (eg, a neutral filter, a light filter) in the delay branch path.

在、J Ϊ t?2°之位置調整為使反射第-和第二探測脈衝 ，光束路徑63 5回到掃描系統6 3 0之時於空間和時間為重 二γ，射第和第一芩考脈衝因而也在沿光束路徑6 3 5”回 ，“統630之時於空間和時間為重疊。由於探測脈衝 /一吊短（約35微微秒），干涉儀延遲支路之長度（延遲路 徑長度）必須製作成與干涉儀受測元件支路長度相等或幾 乎，等使脈衝重疊。在一實施例中，反射鏡72〇、壓電致 動叩725及透鏡715為安裝在一滑動件上容許對延遲路徑長 度做粗略調整。At the position of J Ϊ t? 2 °, the first and second detection pulses are reflected, and the beam path 63 5 returns to the scanning system 6 3 0 when the space and time are double γ. The pulse is therefore also returned along the beam path 6 3 5 "at the time 630, the space and time overlap. Due to the short detection pulses (about 35 picoseconds), the length of the delay branch of the interferometer (the length of the delay path) must be made equal to or almost the same as the length of the branch of the interferometer device under test, so that the pulses overlap. In one embodiment, the mirror 72o, the piezoelectric actuator 725, and the lens 715 are mounted on a slider to allow rough adjustment of the delay path length.

再次參照第一圖，反射脈衝沿光束路徑6 3 5回到光束 轉向光學件634，在此將反射脈衝轉至偏振光束分流器 6 3 2。具備與原始線性偏振正交之線性偏振的反射脈衝由 偏振光束分流器632轉至反射鏡64 2且在此將反射脈衝反射 至光纖耦接器644。光纖耦接器644將反射脈衝輕合於多模 光纖646，此光纖將反射脈衝導引至光纖耦接器648。光纖 輕接器6 4 8將反射脈衝耦合於信號處理次系統6 5 2内之反射 光束光偵測器6 5 0。Referring again to the first figure, the reflected pulse returns to the beam along the beam path 6 3 5 and turns to the optical member 634, where the reflected pulse is transferred to the polarization beam splitter 6 3 2. The reflection pulse having a linear polarization orthogonal to the original linear polarization is transferred from the polarization beam splitter 632 to the mirror 64 2 and the reflection pulse is reflected to the fiber coupler 644 here. The fiber coupler 644 light-couples the reflected pulse to the multimode fiber 646, which guides the reflected pulse to the fiber coupler 648. Optical fiber light connector 6 4 8 couples the reflected pulses to the reflected light in the signal processing subsystem 6 5 2 and the light beam detector 6 5 0.

第15頁 571108 五、發明說明（11) 、於空間和時間重疊之反射第一和第二探測脈衝在光偵 測器6 5 0内產生一干涉電氣信號，此信號之強度由兩道反 射採測脈衝間之光相位差決定。相似地，反射第一和第二 參考脈衝亦在光偵測器65〇内產生一干涉電氣信號，此信 號，強度由兩重疊反射參考脈衝間之光相位差決定。光偵 測器6 5 0為每一對反射探測脈衝及每一對反射參考脈衝在 互阻抗放大器6 5 4之輸入端輸出一電流脈衝。 、 互阻抗放大器654將來自光偵測器65〇之電流脈衝變換 為輸出電壓脈衝，其中探測脈衝干涉信號之輸出電壓脈衝 由快動開關6 5 6耦接於能量分析器6 58而參考脈衝干涉信號 之輸出電壓脈衝由快動開關656耦接於能量分析器66〇。探 測光脈衝和參考光脈衝之輸出電壓脈衝（其持續時間通常 因互阻抗放大器6 54之有限帶寬而為6〇〇毫微秒）在時間方 面充刀隔開使一者可相互區別。一輸出電壓脈衝之電壓_ 時間累計為相應干涉信號能量之度量。能量分析器6 58和 660依慣例各自含有一累計器和一類比數位轉換器，藉此 分別輸出探測干涉信號能量及參考干涉信號能量之丨4位 數位表現。 每一入射探測脈衝和參考脈衝自光纖6 2 6及光纖耦接 = 628入射於掃描系統63〇上之一部分由偏振光束分流器 “專至反射鏡662，在此將脈衝導向光纖耦接器6 64。光 纖耦接器664將該等脈衝耦合於多模光纖6 6 6，此光纖將月 衝導引至光纖耦接器6 68。光纖耦接器6 6 8將入射脈衝耦> 於信號處理次系統65 2内之入射光束光偵測器67〇。 1Page 15 571108 V. Description of the invention (11) Reflections overlapping in space and time The first and second detection pulses generate an interference electrical signal in the photodetector 6 50. The intensity of this signal is obtained by two reflections. The optical phase difference between the measurement pulses is determined. Similarly, reflecting the first and second reference pulses also generates an interference electrical signal in the photodetector 65. The intensity of this signal is determined by the optical phase difference between the two overlapping reflected reference pulses. The optical detector 650 outputs a current pulse at the input of the transimpedance amplifier 654 for each pair of reflected detection pulses and each pair of reflected reference pulses. The transimpedance amplifier 654 converts the current pulse from the photodetector 65 to an output voltage pulse. The output voltage pulse of the detection pulse interference signal is coupled to the energy analyzer 6 58 by the quick-action switch 6 and the reference pulse interferes. The output voltage pulse of the signal is coupled to the energy analyzer 66 by the snap-action switch 656. The output voltage pulses of the detection light pulse and the reference light pulse (the duration of which is usually 600 nanoseconds due to the limited bandwidth of the transimpedance amplifier 6 54) are separated in time by a knife gap. The voltage_time of an output voltage pulse is a measure of the energy of the corresponding interference signal. The energy analyzers 6 58 and 660 conventionally include a totalizer and an analog digital converter, respectively, thereby outputting a 4-digit digital representation of the detected interference signal energy and the reference interference signal energy, respectively. Each incident detection pulse and reference pulse is from the fiber 6 2 6 and the fiber coupling = 628. A part of the incident on the scanning system 63 is polarized beam splitter "specific to the mirror 662, where the pulse is directed to the fiber coupler 6 64. The fiber coupler 664 couples these pulses to the multimode fiber 6 6 6, which guides the moon to the fiber coupler 6 68. The fiber coupler 6 6 8 couples the incident pulses to the signal Incident beam light detector 67 in the processing sub-system 65 2 1

571108571108

光偵測器670為其所偵測之每一入射光 放大器6: 2Λ輸入端輸出-電流脈•。互阻抗放ί i二將 來自光偵/則為6 70之電流脈衝變換為輸出電麼’ 入射探測脈衝信號之輸出電壓二二 能量分析器676而入射來考财^衝：Λ動山開關674耗接於 ^ , 町，考脈衝化號之輸出電壓脈衝由快 ΐ,Γ白人接於能量分析器678。能量分析器676和678依 ::累計器和—類比數位轉換器，I分別輸出 所偵測入身"米測脈衝能量及參考„能量之14位元數位表 現0 在貝知例中’光偵測器6 5 0和6 70為Fermi on ics _The photodetector 670 outputs a current pulse for each incident light amplifier 6: 2Λ input terminal it detects. The transimpedance amplifier ίi converts the current pulse from the light detection / then 6 70 into the output power? The output voltage of the incident detection pulse signal 222 energy analyzer 676 and the incident to test the wealth ^ Chong: Λ 动 山 开关 674 The output voltage pulses consumed by ^, ^, and pulsation number are connected to the energy analyzer 678 by Kuai, Γ white people. Energy analyzers 676 and 678 are based on :: accumulators and analog digital converters. I output the detected pulse energy and the reference 14-bit digital performance of the energy. 0 Detectors 6 5 0 and 6 70 are Fermi on ics _

Corporation 型號FD3 0 0 銦鎵砷化物（inGaAs ) piN 光電二 極體，且互阻抗放大器654和672、快動開關65 6和6 74以及 脈衝此里分析器6 5 8、6 6 0、6 7 6和β 7 8揭示於惠雪等人之美 國專利第5 9 0 5 5 7 7號中。 > 1時信號產生器6 8 0控制雙脈衝光干涉儀系統6〇〇之作 業什％。定時信號產生器6 8 〇耦接於測試向量源6 8 2、鎖模 、 雷射602、光調變器60 8、光調變器616、信號處理次系統 6 52、干涉儀6 9 9以及資料處理控制裝置6 9 7。在一實施例 中’定時信號產生器6 8 0為惠雪等人之美國專利第5 9 〇 5 5 7 7 唬中所揭示雙雷射探測系統之定時信號產生控制的定時信_ 號產生器。 。資料處理控制裝置6 9 7通常為一可程式規劃通用型數 位電腦處理器具備一視訊顯示端子以及習知高速資料攫取 及數位信號處理電路板。Corporation Model FD3 0 0 InGaAs piN Photodiodes with transimpedance amplifiers 654 and 672, snap action switches 65 6 and 6 74 and pulses here analyzers 5 5 8, 6 6 0, 6 7 6 and β 7 8 are disclosed in Hui Xue et al., U.S. Patent No. 5,990,575. > The 1-hour signal generator 680 controls the operation of the dual-pulse optical interferometer system 600%. The timing signal generator 6 8 〇 is coupled to the test vector source 6 8 2, mode-locked, laser 602, light modulator 60 8, light modulator 616, signal processing subsystem 6 52, interferometer 6 9 9 and Data processing control device 6 9 7. In one embodiment, the 'timing signal generator 6 80' is a timing signal generator for controlling the timing signal generation of the dual laser detection system disclosed in U.S. Patent No. 5 095 5 7 7 . . The data processing control device 6 9 7 is usually a programmable digital computer processor with a video display terminal and a conventional high-speed data acquisition and digital signal processing circuit board.

571108 五、發明說明（13) 在一貫施例中，測試向量源68 2 aSchlumberger I TS 9 0 0 0邏輯測試系統。其他測試向量源如簡單資料產 器亦可使用，前提為其時鐘信號夠準確穩定。測試向 682重稷地在多重線684上對受測元件64〇輸出一連串 ^-測試圖）。測試圖連續地循環。也就是說，、當: ::置源682發出測試圖之最終向量，其繞回起始向量且 毫ς開始發出測試圖。測試圖之長度通常為約數微秒至數 ^只鉍例中，鎖模雷射脈衝序列6 0 6及測試向量源 ^射Λ之圖循環為相位鎖定。當#，此並非指個別 相位而是指有關於週期性測試圖循環相位之 物循環週期内有整數的鎖模雷射脈衝： ;求：=:環射脈衝重複率而言，其 一 自衣之長度為1 〇耄微米週期之整數。 接收=ΐ ^二/^^ t咖自測試向量細2 正確頻率時鐘i 各相位鎖定循環技術導出一 環。替代方法^以將鎖I模雷射60 2驅使至測試圖循 供應時鐘信號689且對=虎向產旦生;對鎖模雷射602產生 6 8 8。 、°式向里，原6 8 2產生供應時鐘信號 上輸出1二:：：在的:試圖循環之每-循環開頭在線686 信號產生器_v之—定=開始信號係用來開始定時 疋t序列。在接收到線6 8 6上開始信 第18頁 571108 五、發明說明（14) 號之後一預定時間，定時信號產生器68〇在線6 93上對光 變器6 1 6發出一控制信號。在接收到線6 9 3上控制信號時， 光調變器616藉由容許來自連續波雷射612之連續振幅°光 614通過約150毫微秒定義參考光脈衝618。相似地，在 收到線68 6上開始信號之後一預定時間，定時信號產生器 6 8 0在線6 95上對光調變器6 0 8發出一控制信號。在接收到 線6 9 5上控制信號時，光調變器6〇8藉由容許脈衝序列6〇6 内之一脈衝通過而定義探測光脈衝61〇。又，在接收到 6^6上開始信號之後-預定時間，定時信號產生器_在線 65^ ^ ^ ^6 5 2 # ^ ^ ^ ^ ^ ^ ^ ^ ^ 1 56和674適切地回應切換以便將電壓脈衝從互阻抗放大器 =4和6 72一之輸出端導向脈衝能量分析器6 58、66〇、和 6 士78/如前所述。線6 8 9、691、693和6 9 5上控制信號之 日守係由貧料處理控制裂置6 9 7 * ) 式）設定且以線69〇 ± f t由使用者編寫程 終。 &制L唬與疋時信號產生器6 8 0聯 來自4吕號處理次糸纪g ς q 嫩資料處理控；；=之四個14位元數位輪出信號 而古，資料$视& ΐ衣置因此就測試圖之每一循環 號能量和參考干沣二=和入射芩考脈衝能量和探測干涉信 ^ 〜15號能量。 在參考脈衝維ϋ Λ 位置且探測脈衝以等：：：J686上開始信號在-定定時 移經一系列固定定栌"' 樣方式相對於線686上開始信號 T仇置之時，資料處理控制系統697藉571108 V. Description of the invention (13) In a consistent embodiment, the test vector source 68 2 a Schlumberger I TS 9 0 0 0 logic test system. Other test vector sources such as simple data generators can also be used, provided that their clock signals are accurate and stable. The test outputs a series of ^ -test patterns to the component under test 64 on the multiple line 684. The test pattern is continuously looped. That is, when ::: source 682 sends the final vector of the test pattern, it wraps around the starting vector and starts sending the test pattern. The length of the test pattern is usually from a few microseconds to several milliseconds. In the case of bismuth, the pattern cycle of the mode-locked laser pulse sequence 6 06 and the test vector source ^ radiation Λ is phase locked. When #, this does not refer to individual phases, but refers to objects that have periodic integers in the cyclic phase of the periodic test pattern. There are integer mode-locked laser pulses in the cyclic period:; The length is an integer of a period of 10 μm. Receive = ΐ 二 / / ^ ^ t self-test vector fine 2 correct frequency clock i each phase locked loop technology derived a loop. The alternative method is to drive the mode-locked laser 60 2 to the test chart by supplying a clock signal 689 and to produce a pair of tigers; to the mode-locked laser 602 6 8 8 is generated. ， ° Inward, the original 6 8 2 generates the supply clock signal and outputs 12 :::::: Attempt each cycle-the beginning of the cycle is on line 686 signal generator_v 之 — 定 = Start signal is used to start timing 开始t sequence. After receiving the letter on line 6 8 6 page 18 571108 V. A predetermined time after the invention description (14), the timing signal generator 68 0 sends a control signal to the optical converter 6 1 6 on line 6 93. Upon receiving the control signal on line 6 9 3, the light modulator 616 defines a reference light pulse 618 by allowing continuous amplitude light 614 from the continuous wave laser 612 to pass through approximately 150 nanoseconds. Similarly, a predetermined time after receiving the start signal on line 68 6, the timing signal generator 6 8 0 sends a control signal to the light modulator 6 0 8 on line 6 95. Upon receiving the control signal on line 6 95, the optical modulator 608 defines the detection light pulse 61 by allowing one of the pulses in the pulse sequence 606 to pass. Also, after receiving the start signal on 6 ^ 6-a predetermined time, the timing signal generator _ online 65 ^ ^ ^ ^ 6 5 2 # ^ ^ ^ ^ ^ ^ ^ ^ ^ 1 56 and 674 respond appropriately to the switch in order to The voltage pulses are directed from the outputs of the transimpedance amplifiers = 4 and 6 to the pulse energy analyzers 6 58, 66, and 6 ± 78 / as previously described. The day guards of the control signals on lines 6 8 9, 691, 693, and 6 9 5 are set by lean processing control and split (6 7 7 *)) and are programmed by the user with a line of 690 ± f t. & System L8 and Time signal generator 6 8 0 from 4 Lu No. processing time g ς q tender data processing control; = four 14-bit digital rotation signal and ancient, data $ 视 & amp Therefore, the kimono sets the energy of each cycle of the test chart and the reference energy = and the energy of the incident test pulse and the energy of the detection interference signal ^ ~ 15. At the reference pulse dimension ϋ Λ position and the detection pulse is equal to ::: J686, the data is processed when the start signal shifts through a series of fixed settings at a fixed timing " 'relative to the start signal T on line 686. Control system 697

57H〇857H〇8

五、發明說明（15) 由自四個脈衝能量分析器收集資料攫取或 一 氣信號。一般而言使用5 0 〇個不同定時位’貝70件6 4 〇上電 定時位置計算探測反射比與參考反射比 。為每一探測 代表受測元件640上與測試圖波形相應之比厭(列树，且理解為 在一實施例中，施行通過定時位置 i ’交化。 -迭代。在此迭代過程中，探測脈衝η之探測脈衝單 波形雜訊減少。在另一實施例；貝處理使 10次至約1〇〇。次迭代。在此等迭代期間，探:列之約 至約100個測試圖循環保持在每一定時位置。/、、仿對約10 /光調變器60 8並不完全擋住鎖模脈衝序列6〇6内之 脈衝。殘餘鎖模脈衝漏出光調變器6〇8，且在光 和670内造成偏置信號。在一實施例中，使用$ :箄： 美國專利第5 90 5 577號所揭示之一種方法藉以；偏置产 號強度並自探測干涉信號和參考干涉信號中減去。。 探測干涉信號和參考干涉信號之強度對於干涉儀699 内受測元件支路與延遲支路間之光學路徑長度差異很敏 感。因此，該等干涉信號對於由入射探測脈衝和參考脈衝 建立之欠測元件640沿光學軸線位置报敏感。受測元件64〇 沿光學軸線之位置在下文稱為受測元件6 4 〇之Z位置。 干涉^號對受測元件6 4 0 Z位置之敏感度由下述實例顯 明。考慮一道自受測元件64 0反射之參考脈衝（亦即反射 第一參考脈衝）且稱之為光束A。同樣考慮一道自干涉儀 延遲支路反射鏡7 2 0反射之參考脈衝（亦即反射第二參考V. Description of the invention (15) Data are collected from four pulse energy analyzers or a gas signal. Generally speaking, 50,000 different timing bits are used to calculate the detection reflectance and the reference reflectance at 70 timings and 64 positions. For each probe, the ratio corresponding to the waveform of the test pattern on the device under test 640 (column tree, and understood to be an intersection through the timing position i 'in one embodiment.-Iteration. During this iteration, the probe The detection pulse single-wave noise of the pulse η is reduced. In another embodiment, the bead processing makes 10 to about 100. iterations. During these iterations, the number of probes from about to about 100 test patterns is maintained cyclically. At each timing position. / ,, the pseudo-pair about 10 / the light modulator 60 8 does not completely block the pulses in the mode-locked pulse sequence 606. The residual mode-locked pulse leaks out of the light modulator 6008, and The light and 670 cause a bias signal. In one embodiment, $: 箄 is used: a method disclosed by US Patent No. 5 90 5 577; the strength of the production number is biased and the interference signal and the reference interference signal are self-detected Subtract ... The intensity of the detection interference signal and the reference interference signal is sensitive to the difference in the optical path length between the branch of the component under test and the delay branch in the interferometer 699. Therefore, these interference signals are sensitive to the incident detection pulse and the reference pulse. Established under test element 6 The position of 40 along the optical axis is reported as sensitive. The position of the tested element 64o along the optical axis is hereinafter referred to as the Z position of the tested element 6 4 0. The sensitivity of the interference ^ sign to the position of the tested element 6 4 0 Z is given by The example is clear. Consider a reference pulse reflected from the measured element 64 0 (that is, the first reference pulse is reflected) and call it beam A. Also consider a reference pulse reflected from the interferometer delay branch mirror 7 2 0 ( Second reference

第20頁 571108 五、發明說明（16) 脈衝）且稱之為光束B。假設反射鏡72〇之位置調整 射第一和第二參考脈衝對在光偵測器6 5 〇之時間和空 疊度最大化。假設光束A之功率不隨受測元件64〇z位置改 ί 64〇\1 ί^ ^ "72° ^ ^ ^ ^ ^ ^ ^ ^ 件^之2位置移動四分之—光波長之距離κ貞測器_ 所測付之干涉信號總功率會在A和6之功率和附近變動。由 基f光學理論可得到在此z位置範圍内發現之最大和最小 功率（相應於建設性干涉條件和破壞性 ^^:)2和（^|2)2。平均功率為“/:例刀來別為 =瓦倘功率為1微瓦（Μ )且光束k功率為3 η 536微瓦，平均測得功率為4微瓦。假定光脈衝2波長為 64毫微米，光偵測器輸出隨Ζ位置之變動在第：圖中：’” f弦曲線83G顯示。圖中可見(Μ微米之Ζ位圖= 相應於受測元件64〇之一丨把1直又化（例如 大變化。 牛4〇之一小振幅振動）造成測得功率有極 位置二二議受測元侧 51, ^ 生猎由k供反射鏡720位置之閉路控制得 =偵在受™:二 當作支撐喜a . 捉仏昧動致動杰（例如 之運動在源頭,分）且控制該致動器將受測元件64〇 此二實ίβ處抵銷。以下控制裝置及方法的說明應用於 第21頁 571108 五、發明說明（17) ___ ★鐘m:圖和第二圖，藉由探測光束或參考光击 ί轉向=件634光柵掃描受測元件64。為受判光 像，此時快門703為關閉以避、1凡件640造 然後入射束探測脈衝和夹_ 可干涉圖案。 對正受測元件640上之—相考二，由光束轉向光學件634Page 20 571108 V. Description of the invention (16) Pulse) and it is called beam B. Assume that the position of the mirror 72o is adjusted to maximize the time and overlap of the first and second reference pulse pairs at the photodetector 6500. Assume that the power of beam A does not change with the position of the tested element 64〇z. 64〇 \ 1 ^ ^ ^ 72 ° ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ 2 of the position ^ is moved by a quarter of a distance of the light wavelength κ Chastity detector_ The total power of the interference signal measured will vary between the power of A and 6. The maximum and minimum powers (corresponding to constructive interference conditions and destructive ^^ :) 2 and (^ | 2) 2 found in this z-position range can be obtained from the fundamental f optical theory. The average power is "/: For example, if the power is 1 watt, if the power is 1 microwatt (M) and the beam k power is 3 η 536 microwatts, the average measured power is 4 microwatts. Assume that the light pulse 2 has a wavelength of 64 milliseconds. Micron, the change of the photodetector output with the Z position is shown in the figure: '' f chord curve 83G is displayed. It can be seen in the figure (Z-bit map of M micron = corresponding to one of 64 of the tested element. Straightening 1 (for example, large change. Small amplitude vibration of 40%) causes the measured power to have a pole position. The side of the tested element 51, ^ The hunting is controlled by the closed circuit of the position of the mirror 720 for k = detection in the receiving ™: the two are used to support the hi a. Catching ambiguous actions (such as the movement at the source, points) and Controlling the actuator offsets the test component 64 ° from the two real β. The following description of the control device and method is applied on page 21 571108 V. Description of the invention (17) ___ ★ Zhong m: Figure and the second figure, The detection beam or reference light is used to turn and turn = 634 raster scans the test element 64. For the light image to be judged, the shutter 703 is closed at this time to avoid the 640 and then the incident beam detection pulse and clip_ can interfere Alignment on the component under test 640—Consider the second, the beam is turned to the optical element 634

為開放。資料處理控制Λ 且快門7〇3和開關WO 器（mc) 745送出==69\在_上對數位類比轉換 W在其輸出端產生心， 流。放大議在其輪n而2阻議内產生-電 和790之電流她和成卜7 舁，、輸入鈿來自電阻器785 接於壓電致動器725，兮致勒叫、έ ^ 輸出電壓耦 725對約100伏特之放士^動益連結於反射鏡MO。致動器 動大約6微米。、 大為735輸出電壓變化將反射鏡720移 在開關7 4 0為開路日卑，、s σσ 大器735之輪出产获Λl匕電阻器790之電流為零且放 制。資料卢採批。J虎文位類比轉換器745之輸出信號控 設定-系二電‘ ^ Ϊ置697在數位類比轉換器745之輸出端 變延遲支路之路彳a ; Ϊ致動器725移動反射鏡720並因而改 數位輸出表現的；g涉=錄由脈衝能量分析器66°之 範圍内之—反：6二;$任何Z位置而言’經發現在-半波長 干涉信號功率範=位置會Λ光债測器650給出一在參考 實例中， …中央之測知麥考干涉信號功率。在上述 藉由資料卢力率、及為4微瓦。一個代表此中點功率之電壓 处理控制裝置6 9 7設定在數位類比轉換器7 7 5之輸For opening. The data processing control Λ and the shutter 703 and the switch WO device (mc) 745 send == 69 \ on the _ analog logarithmic conversion W generates a heart, flow at its output. Amplify the current generated in its round n and 2-the current of electricity and 790, and 卜 7 舁, the input 钿 comes from the resistor 785 connected to the piezoelectric actuator 725, so the yelling, ^ ^ output voltage A couple of 725 pairs of approximately 100 volts are connected to the mirror MO. The actuator moves about 6 microns. The change in the output voltage of 735 shifts the mirror 720 at the switch 7 40 for the open circuit. The current of the s σσ resistor 735 produced by the Δl resistor 790 is zero and controlled. Information Lu Cai approved. J Tiger text bit analog converter 745 output signal control settings-Department of Electricity '^ Set 697 at the output end of the digital analog converter 745 variable delay branch road 之 a; Ϊ actuator 725 moves the mirror 720 and Therefore, the digital output is changed; g == recorded by the pulse energy analyzer within the range of 66 °-inverse: 6 2; $ at any Z position 'is found in the half-wavelength interference signal power range = position will be Λ light The debt detector 650 gives a reference example, the central measurement of the McCaw interference signal power. In the above, the data rate is 4 microwatts. A voltage representing the power at this midpoint. The processing control device 6 9 7 is set to the output of the digital analog converter 7 7 5

571108 五、發明說明（18) 出端上且用作對一兽公访士哭ry ^ Λ 徐处曰八折哭放大态77〇之麥考輸入。一個由脈 衝旎置分析窃6 6 0產生日讲本4 i τ μ ®柞f+i ^ 翏考干涉信號強度之類比信 就b b 1用作對差分放大哭7 7 π +。 . Λ maQ7 ^ ^ ^ ^ σ〇 770之另一輸入。資料處理控制裝 置6 9 7將數位類比鍊拖哭7 /1 r七μ丨 ν 轉換的7 4 5之輪出信號設定於中階，豆 放大器735之輸出設定為約5〇伏特。 ，、肘 f料處理控制裝置697使開關7 40閉路讓一電流從累計 放大器75 0之輸出端經過電阻器79〇流到放大器73 5之輸入 端。參考數位類比轉換器77 5輸出信號與脈衝能量分析器 6 6 0之類&比輸出仏號6 6 !之間的差異顯現在差分放大器口 〇 之輸出端。定時信號產生器6 8 〇在線6 9 3上對單穩態電路 7 8 0發出控制彳§號，且如前所述對光調變器6丨6發出信號。 單穩恶電路780以約比1〇〇微秒長的時間間隔將開關765閉 路約5微秒。若差分放大器77〇之輸出信號在開關765為閉 路時並非零伏特，一電流通過電阻器7 6 〇流到累計放大器 750 ’改變放大器750輸出端之電壓及通過電阻器790之電 流’且經由放大器735和壓電致動器725改變反射鏡720的 位置。 放大器7 7 0、7 5 0和7 3 5構成一個控制反射鏡7 2 0位置之 負反饋迴路。當受測元件6 4 0之Z位置改變，該負反饋迴路 改變反射鏡7 2 0位置以將測得參考干涉信號維持在其功率 範圍中央（在上述實例中為4微瓦）。反射鏡720的質量為 小，是以該反饋迴路能夠在約2 0 0赫茲或更大之頻率範圍 DC依循受測元件6 4 0之振動。反射鏡7 2 0的位置能由壓電裝 置7 2 5或一移動線圈電磁裝置。如先前所述，此反饋可施571108 V. Description of the invention (18) It is used at the beginning and is used as the McCao input for crying ry ^ Λ Xu Chuyu's 20% off crying magnified state. An analogue letter from the pulse analysis analysis of 6 6 0 is 4 i τ μ ® 柞 f + i ^ Consider the analog signal strength of the interference signal. Let b b 1 be used as a differential amplification cry 7 7 π +. Λ maQ7 ^ ^ ^ ^ σ〇 770 is another input. The data processing control device 6 9 7 sets the digital analog chain tow 7/1 r seven μ 丨 ν converted 7 4 5 wheel output signal to the intermediate stage, the output of the bean amplifier 735 is set to about 50 volts. The elbow f material processing control device 697 closes the switch 7 40 to allow a current to flow from the output terminal of the accumulation amplifier 75 0 through the resistor 79 0 to the input terminal of the amplifier 73 5. The difference between the reference digital analog converter 77 5 output signal and the pulse energy analyzer 6 6 0 & specific output number 6 6! Appears at the output of the differential amplifier port 〇. The timing signal generator 6 8 0 sends a control signal to the monostable circuit 7 8 0 on the line 6 9 3 and sends a signal to the light modulator 6 6 as described above. The monostable wick circuit 780 closes the switch 765 for about 5 microseconds at intervals longer than about 100 microseconds. If the output signal of the differential amplifier 77 ° is not zero volts when the switch 765 is closed, a current flows through the resistor 76 to the accumulation amplifier 750 'changes the voltage at the output of the amplifier 750 and the current through the resistor 790' and passes through the amplifier 735 and piezoelectric actuator 725 change the position of the mirror 720. The amplifiers 7 7 0, 7 50, and 7 3 5 form a negative feedback loop that controls the position of the mirror 7 2 0. When the Z position of the device under test 640 changes, the negative feedback loop changes the position of the mirror 720 to maintain the measured reference interference signal in the center of its power range (4 microwatts in the above example). The mass of the reflecting mirror 720 is small, so that the feedback loop can follow the vibration of the device under test 640 in a frequency range of about 200 Hz or greater. The position of the reflecting mirror 7 2 0 can be made by a piezoelectric device 7 2 5 or a moving coil electromagnetic device. As mentioned earlier, this feedback can be applied

571108571108

=於雙測元件64 0處之一類似致動器以在振動源頭將其抵 ^，不過适種方式比較困難，因為受測元件64〇通常比 射鏡720大得多。 久= One of the similar actuators at the dual test element 64 0 to counteract it at the source of vibration, but this method is more difficult, because the test element 64 is usually much larger than the mirror 720. Long

以干涉儀叉測兀件支路和干涉儀延遲支路之光學路猝 長度失配的函數表示之測得參考干涉信號功率亦在第三= 曲線圖中以曲線830表現。由數位類比轉換器口5產生之: 考級由線805表示。圖中可見在反饋迴路有作用的條件多 下、’ ^數個穩定（負反饋）和不穩定（正反饋）反射鏡伋 置為苓考干涉彳s號功率63〇等於參考級8〇5。這些位置發 在點810、82 0、840和8 5 0。在受測元件640沒有振動的條 件下，反射鏡72 0之位置會受驅使離開不穩定點朝穩定點 f Ϊ之一移動，最後停留在非常接近穩定點其中之一處。 付，供電路系統將反饋迴路内電子增益反相使穩定點和不 ，定點的位置顛倒。重點在於反饋系統夠有效能夠在一般 文測π件振動存在時維持接近其穩定點之作業，否則反^ 系統會一直尋找一新的穩定作業點。如業界所熟知，反饋 之有效性有部分係由放大器7 3 5、7 5 0和7 7 0之增益和 f見’與反射鏡720連結之壓電致動器72 5的響應時間，以 及開關76 5閉路對放大器770輸出及類比信號661取樣之頻The measured reference interference signal power expressed as a function of the optical path burst length mismatch between the interferometer fork measuring element branch and the interferometer delay branch is also represented by the curve 830 in the third graph. Generated by digital analog converter port 5: The grade is indicated by line 805. It can be seen in the figure that under the conditions where the feedback loop is effective, several stable (negative feedback) and unstable (positive feedback) mirrors are set to Lingkao Interference 彳 s power 63 ° equal to the reference level 805. These positions occur at points 810, 82 0, 840, and 8 50. Under the condition that the measured element 640 is not vibrating, the position of the mirror 72 0 is driven to move away from the unstable point toward one of the stable points f ，, and finally stays very close to one of the stable points. Pay, for the circuit system to reverse the electronic gain in the feedback loop to make the stable point and not, the position of the fixed point is reversed. The important point is that the feedback system is effective enough to maintain the operation close to its stable point in the presence of general vibration of the measured π piece, otherwise the inverse system will always look for a new stable operating point. As is well known in the industry, the effectiveness of feedback is partly due to the gain and f of amplifiers 7 35, 7 50 and 7 70. See the response time of the piezoelectric actuator 72 5 connected to the mirror 720 and the switch 76 5 Closed circuit sampling frequency of amplifier 770 and analog signal 661

率決定。 旦 在一實施例中，定時信號產生器68〇修改為對測試向 夏=682所輸出測試圖之每一循環在線6 9 3上提供一控制信 號起過久。因此在每一測試圖循環中，參考脈衝6 1 8定 義及開關76 5閉路超過一次，藉此提高反饋系統對類比信Rate decision. In one embodiment, the timing signal generator 68 is modified to provide a control signal on line 6 9 3 for each cycle of the test pattern output from the test direction Xia = 682. Therefore, in each test pattern cycle, the reference pulse 6 1 8 is defined and the switch 76 5 is closed more than once, thereby improving the analog signal of the feedback system.

第24頁 571108 I — 五、發明說明（20) 號6 6 1之取樣率 探測雷射與參考雷射間之任何 涉圖和麥考干涉圖不同。舉 波長差異導致探挪干 長比參考雷射短’則以受測元件：路：：探剛雷射之光波 徑長度失配的函數表示之測得 ：延遲支路之光學路 曲線830相似之一正弦曲線表現/、/為W虎可用與第三圖 就夠大之光學路徑長度失配而+，1—π為較短波長。因此， 元件640運動造成的變動不5 ’振^干涉信號内因受測 鏡720移動而得到補償。舉例=疋芬考干涉信號之反射 波長雷射相差約〇. 5毫微米，;屯他若標稱1 064毫微米 之光學路徑長度失配量應約小於^Ο、1 70件支路與延遲支路 倘右使用一共焦顯微鏡將探 /、 、 受測元件640，則受測元件64$ ^ j脈衝和參考脈衝傳送到 幅及其光學相位。因此，會 j動能改變反射脈衝之振 償反射光之振幅變化而非反仃些反射鏡7 2 0移動以補 件640係在物鏡636之焦點，予Ϊ:立變化。由於受剛元 常低於相關聚焦範圍± ! 〇 % )、=此寻振幅變化為夠小（通 而此效應得以忽略。 〇 會由夠小的反射鏡移動補償 入射麥考脈衝之振幅變 錯誤補償運動。在一餘# / 相似地會導致反射鏡720之 員％例中 + $ 反射比而非參考脈衝干哚 匕寺‘誤藉由以麥考脈衝 過，參考脈衝振幅變動驅7反饋迴路得以避免。不 反射鏡72 0錯誤移動引起之\吳 '③於5% ’因而小到足以忽略由 本說明書僅為範例日τ '日吟。 〃、限制意義。習於此技藝者經 第25頁 571108Page 24 571108 I — V. Sampling rate of invention note (20) No. 6 6 1 Any detection pattern and reference interference pattern between the detection laser and the reference laser are different. The difference in wavelength causes the probe length to be shorter than the reference laser 'is measured as a function of the measured component: path :: optical path length mismatch of the stiff laser, measured: the optical path curve 830 of the delay branch is similar A sinusoidal curve represents /, / is the optical path length mismatch between W tiger and the third figure is large enough, +, 1-π is a shorter wavelength. Therefore, the variation caused by the movement of the element 640 is not compensated by the movement of the test lens 720 within the 5 'vibration interference signal. Example = The laser wavelength difference of the reflection wavelength of the Fencao interference signal is about 0.5 nanometers; if the nominal optical path length mismatch of Tunta is 1 064 nanometers, it should be less than ^ 0, 1 70 branches and delay If a confocal microscope is used in the branch circuit to detect, test, and test the element 640, the test element and the reference pulse are transmitted to the amplitude and its optical phase. Therefore, the kinetic energy of the reflected pulse can be changed to compensate for the change in the amplitude of the reflected light instead of reflecting the mirror 7 2 0 to move the patch 640 to the focal point of the objective lens 636, which can cause a vertical change. Because the receiving element is usually lower than the relevant focus range ±! 〇%), = the amplitude variation is small enough (this effect is ignored). 〇 The amplitude of the incident McCaw pulse will be compensated by the small enough mirror movement. Compensating motion. In a case where ## / similarly would cause a mirror 720 member% + reflection ratio instead of a reference pulse, the dry pulse is incorrectly driven by the McCaw pulse, and the reference pulse amplitude drives the 7 feedback loop. Avoid it. The non-mirror 72 0 caused by the wrong movement of \ Wu'③ at 5% 'is therefore small enough to ignore the example of this manual, only τ' Riyin. 〃, restrictive meaning. Learned by this artist on page 25 571108

第26頁 571108 圖式簡單說明 6 0 0 :雙脈 602 604 606 608 609 610 612 614 618 620 622 623 624 626 鎖模 雷射 輸出 616 探測 探測 連續 連續 參考 光束 642 偏振 628 偏振 6 3 0 ··掃描 632 ^ 701 634 635 636 638 639 640 光束 光束 探測 活性 710 受測 衝光干涉儀系統 雷射源 平台 雷射脈衝序列 光調變器 脈衝 光脈衝 波雷射源 振幅雷射束 脈衝 組合器 662 > 705 ^ 720 器 644 ^ 648 > 664 單模光纖 系統 光束分流 轉向光學件 脈衝成形光學件 四分之一波片 元件 反射鏡 光纖耦I接器 透鏡） 641 : x-y 臺Page 26 571108 Brief description of the diagram 6 0 0: Double pulse 602 604 606 608 609 610 612 614 618 620 622 623 624 626 Mode-locked laser output 616 Detection continuous continuous reference beam 642 Polarization 628 Polarization 6 3 0 ·· Scan 632 ^ 701 634 635 636 638 639 640 Beam Detector activity 710 Measured laser interferometer system Laser source platform Laser pulse sequence Light modulator Pulse light Pulse wave Laser source Amplitude Laser pulse combiner 662 > 705 ^ 720 644 ^ 648 > 664 single mode fiber system beam splitting steering optics pulse forming optics quarter-wave plate element reflector fiber optic coupler lens 641: xy stage

第27頁 571108 圖式簡單說明 646、6 68 :多模光纖 6 5 0 :反射光束光偵測器 6 5 2 :信號處理次系統 6 54、6 72、73 5、75 0、770 :放大器 6 5 6、6 7 4 :輸出電壓脈衝由快動開關 6 5 8、6 6 0、6 7 6、6 7 8 :脈衝能量分析器 6 6 1 :反射連續波信號 6 7 0 :入射光束光偵測器 6 8 0 :定時信號產生器 6 8 2 :測試向量源 684 :多重線 6 8 5 :設定 6 8 6 :開始信號 6 8 8 :時鐘信號 6 8 9 :正確頻率時鐘信號 6 9 3 :取樣 6 9 7 ··資料處理作業控制介面及顯示裝置 6 9 9 :干涉儀及電子控制系統 703 :快門 7 0 4 :光束路徑 7 1 2 :孔徑 715 :透鏡 7 2 5 :壓電致動器 740 、 765 :開關Page 27 571108 Schematic description 646, 6 68: Multimode fiber 6 50: Reflected light beam detector 6 5 2: Signal processing subsystem 6 54, 6 72, 73 5, 75 0, 770: Amplifier 6 5 6, 6 7 4: The output voltage pulse is actuated by the fast-acting switch 6 5 8, 6 6 0, 6 7 6, 6 7 8: Pulse energy analyzer 6 6 1: reflected continuous wave signal 6 7 0: incident beam light detection Tester 6 8 0: Timing signal generator 6 8 2: Test vector source 684: Multi-line 6 8 5: Setting 6 8 6: Start signal 6 8 8: Clock signal 6 8 9: Clock signal with correct frequency 6 9 3: Sampling 6 9 7 ·· Data processing operation control interface and display device 6 9 9: Interferometer and electronic control system 703: Shutter 7 0 4: Beam path 7 1 2: Aperture 715: Lens 7 2 5: Piezo actuator 740, 765: switch

571108571108

第29頁Page 29

Claims (1)

571108 六、申請專利範圍 1 · 一種在重複電測試圖施加於半導體元件時偵測該元件内 活動之方法，其包含以下步驟： 在該測試圖之每一重複中於一選定時間提供一第一探測 光脈衝並分流該第一探測光脈衝以提供至少一第二探測光 脈衝； 在該電測試圖之每一重複中於一相對於提供該第一探測 光脈衝之該選定時間而遷移的時間，提供一第一參考光脈 衝並分流該第一參考光脈衝以提供至少一第二參考光脈 衝； 將該第一探測光脈衝和該第一參考光脈衝導向至該半導 體元件上； 在該第一探測光脈衝和該第一參考光脈衝與該半導體元 件相互作用之後，使該第一探測光脈衝與該第二探測光脈 衝結合且使該第一參考光脈衝與該第二參考光脈衝結合， 而使得該第一探測光脈衝與該第二探測光脈衝在空間及時 間為重疊且該第一參考光脈衝與該第二參考光脈衝在空間 及時間為重疊； 偵測該重疊探測光脈衝，藉以提供一探測干涉信號； 偵測該重疊參考光脈衝，藉以提供一參考干涉信號；以 及 決定該探測干涉信號和該參考干涉信號之一函數。 2.如申請專利範圍第1項之方法，其更包含隨測試圖之重 複改變該所選時間，而使得該探測干涉信號和該參考干涉 信號之該函數係由該測試圖内複數個該選定時間所決定。571108 VI. Scope of patent application1. A method for detecting the activity in a semiconductor device when a repeated electrical test pattern is applied to the semiconductor device, which includes the following steps: Provide a first at a selected time in each repetition of the test pattern Probe light pulses and shunt the first probe light pulses to provide at least one second probe light pulse; a time to migrate with respect to the selected time at which the first probe light pulse is provided in each repetition of the electrical test pattern Providing a first reference light pulse and shunting the first reference light pulse to provide at least a second reference light pulse; directing the first detection light pulse and the first reference light pulse to the semiconductor element; in the first After a detection light pulse and the first reference light pulse interact with the semiconductor element, the first detection light pulse is combined with the second detection light pulse and the first reference light pulse is combined with the second reference light pulse So that the first detection light pulse and the second detection light pulse overlap in space and time and the first reference light pulse and the second reference light pulse The overlap is in space and time; detecting the overlapping detection light pulse to provide a detection interference signal; detecting the overlapping reference light pulse to provide a reference interference signal; and determining the detection interference signal and the reference interference signal. A function. 2. The method according to item 1 of the scope of patent application, further comprising changing the selected time repeatedly with the test pattern, so that the function of the detection interference signal and the reference interference signal is selected by a plurality of the selections in the test pattern. Time is up. 571108 六、申請專利範圍 其中該第一探測光脈衝 3. 如申請專利範圍第1項之方法 為一雷射光脈衝。 其更包含自一鎖模雷射 4. 如申請專利範圍第3項之方法 光脈衝序列選擇該雷射光脈衝。 其更包含相位鎖定該鎖 5. 如申請專利範圍第4項之方法 模雷射光脈衝序列和該重複測試圖。 6 ·如申請專利範圍第1項之方法，其中該第一參考光脈衝 為一雷射光脈衝。 其更包含調變一雷射束 其中該等探測光脈衝和 其更包含在該電測試圖 ，而提供該第一參考光脈 ，其更包含將該第一探測 7. 如申請專利範圍第6項之方法 以提供該雷射光脈衝。 8. 如申請專利範圍第1項之方法 參考光脈衝之波長大致相同。 9 ·如申請專利範圍第1項之方法‘ 之每一重複中將相對於該測試圖 衝之該時間之固定。 1 0.如申請專利範圍第1項之方法 光脈衝和該第一參考光脈衝沿一相同路徑導向至該半導體 元件上。 1 1.如申請專利範圍第1項之方法，其更包含在該第二探測 光脈衝與該第一探測光脈衝結合及該第二參考光脈衝與該 第一參考光脈衝結合之前，沿一延遲路徑導引該第二探測 光脈衝和該第二參考光脈衝。 1 2.如申請專利範圍第1 1項之方法，其更包含反饋控制該 延遲路徑之長度以補償該半導體元件之運動。571108 6. The scope of patent application Where the first detection light pulse 3. The method of item 1 of the scope of patent application is a laser light pulse. It also includes a self-locking laser. 4. The method of item 3 of the scope of patent application The light pulse sequence selects the laser light pulse. It further includes phase locking the lock. 5. The method of the scope of patent application No. 4 The mode laser light pulse sequence and the repeated test pattern. 6. The method of claim 1 in the patent application range, wherein the first reference light pulse is a laser light pulse. It further includes modulating a laser beam in which the detection light pulses and it are further included in the electrical test chart, and providing the first reference light pulse, which further includes the first detection 7. Such as the scope of patent application No. 6 Method to provide the laser light pulse. 8. The method of applying patent No. 1 is about the same wavelength of reference light pulse. 9 · The method of item 1 of the scope of patent application ‘each iteration will be fixed relative to the time of the test pattern. 10. The method according to item 1 of the scope of patent application. The light pulse and the first reference light pulse are directed to the semiconductor element along a same path. 1 1. The method according to item 1 of the patent application scope, further comprising, before the second detection light pulse is combined with the first detection light pulse and the second reference light pulse is combined with the first reference light pulse, a A delay path guides the second detection light pulse and the second reference light pulse. 1 2. The method according to item 11 of the patent application scope, further comprising feedback controlling the length of the delay path to compensate the movement of the semiconductor element. 第31頁 571108 六、申請專利範圍 1 3.如申請專利範圍第1 2項之方法，其中該反饋控制使該 參考干涉信號穩定。 1 4.如申請專利範圍第1項之方法，其更包含放大該探測干 涉信號，累計該探測干涉信號，及數位化該探測干涉信 號。 1 5.如申請專利範圍第1項之方法，其更包含放大該參考干 涉信號，累計該參考干涉信號，及數位化該參考干涉信 號。 1 6.如申請專利範圍第1項之方法，其中偵測該重疊的探測 光脈衝及該偵測重疊的參考光脈衝之步驟包含偵測該重疊 的探測光脈衝及該重疊的參考光脈衝，以提供一偵測器信 號，並將該偵測器信號之一部分累計並數位化以提供一探 測干涉信號，且將該偵測器信號之一部份累計並數位化以 提供一參考干涉信號。 1 7.如申請專利範圍第1項之方法，其更包含消去該探測干 涉信號和該參考干涉信號内之振幅雜訊。 1 8.如申請專利範圍第1 7項之方法，其中消去振幅雜訊之 步驟包含偵測一探測光脈衝能量之一部分，偵測一參考光 脈衝能量之一部分，藉由該部分探測光脈衝能量正常化該 探測干涉信號，及藉由該部分參考光脈衝能量正常化該參 考干涉信號。 1 9. 一種在重複電測試圖施加於半導體元件時偵測該元件 内活動之裝置，其包含： 一第一探測光脈衝源，藉此在該電測試圖之每一重複中Page 31 571108 6. Scope of patent application 1 3. The method according to item 12 of patent scope, wherein the feedback control stabilizes the reference interference signal. 14. The method according to item 1 of the scope of patent application, further comprising amplifying the detection interference signal, accumulating the detection interference signal, and digitizing the detection interference signal. 15. The method according to item 1 of the scope of patent application, further comprising amplifying the reference interference signal, accumulating the reference interference signal, and digitizing the reference interference signal. 16. The method according to item 1 of the scope of patent application, wherein the step of detecting the overlapping detection light pulse and detecting the overlapping reference light pulse includes detecting the overlapping detection light pulse and the overlapping reference light pulse, A detector signal is provided, a part of the detector signal is accumulated and digitized to provide a detection interference signal, and a part of the detector signal is accumulated and digitized to provide a reference interference signal. 1 7. The method according to item 1 of the patent application scope, further comprising eliminating amplitude noise in the detection interference signal and the reference interference signal. 1 8. The method according to item 17 of the scope of patent application, wherein the step of eliminating amplitude noise includes detecting a portion of a detection light pulse energy, detecting a portion of a reference light pulse energy, and detecting the light pulse energy by the portion. Normalize the detection interference signal, and normalize the reference interference signal by the part of the reference light pulse energy. 1 9. A device for detecting a movement in a semiconductor element when a repeated electrical test pattern is applied to the semiconductor element, comprising: a first probe light pulse source, whereby in each repetition of the electrical test pattern 571108 六、申請專利範圍 於一選定時間輸出一第一探測光脈衝； 一第一參考光脈衝源，藉此在一相對於提供該第一探測 光脈衝之該選定時間而遷移的時間，輸出一第一參考光脈 衝； 一分流器，該第一探測光脈衝和該第一參考光脈衝入射 於其上，藉以提供至少一第二探測光脈衝及第二參考光脈 衝； 一支撐器，其支撐該第一探測光脈衝和第一參考光脈衝 入射之半導體元件； 一光束組合器，其定位為在該第一探測光脈衝和該第一 參考光脈衝與該半導體元件相互作用之後，使該第一探測 光脈衝與該第二探測光脈衝結合且使該第一參考光脈衝與 該第二參考光脈衝結合； 一偵測器，其定位為偵測第一探測光脈衝和該第二探測 光脈衝，藉此輸出一探測干涉信號，且偵測該第一參考光 脈衝和該第二參考光脈衝，藉此輸出一參考干涉信號；及 一處理器，其與該偵測器耦接，藉此決定該探測干涉信號 和該參考干涉信號之函數。 2 0.如申請專利範圍第1 9項之裝置，其中該所選時間隨該 電測試圖之該重複而改變，使得該探測干涉信號和該參考 干涉信號之該函數係由該測試圖内複數個該選定時間所決 定。 2 1.如申請專利範圍第1 9項之裝置，其中該第一探測光脈 衝源包含一第一雷射，該第一參考光脈衝源包含一第二雷571108 VI. The scope of the patent application outputs a first detection light pulse at a selected time; a first reference light pulse source, thereby outputting a time that migrates relative to the selected time at which the first detection light pulse is provided; A first reference light pulse; a shunt, on which the first detection light pulse and the first reference light pulse are incident, to provide at least a second detection light pulse and a second reference light pulse; a supporter, which supports The semiconductor element on which the first detection light pulse and the first reference light pulse are incident; a beam combiner positioned to cause the first detection light pulse and the first reference light pulse to interact with the semiconductor element so that the first A detection light pulse is combined with the second detection light pulse and the first reference light pulse is combined with the second reference light pulse; a detector is positioned to detect the first detection light pulse and the second detection light A pulse, thereby outputting a detection interference signal, and detecting the first reference light pulse and the second reference light pulse, thereby outputting a reference interference signal; and a processor , Which is coupled to the detector, thereby determining a function of the detection interference signal and the reference interference signal. 20. The device according to item 19 of the scope of patent application, wherein the selected time is changed with the repetition of the electric test pattern, so that the function of the detection interference signal and the reference interference signal is a complex number in the test pattern. It depends on the selected time. 2 1. The device according to item 19 of the patent application scope, wherein the first detection light pulse source includes a first laser, and the first reference light pulse source includes a second laser. 571108 六、申請專利範圍 射。 2 2.如申請專利範圍第1 9項之裝置，其中該第一探測光脈 衝源與該第一參考光脈衝源包含一相同雷射。 2 3.如申請專利範圍第1 9項之裝置，其中該第一探測光脈 衝源為一雷射。 2 4.如申請專利範圍第23項之裝置，其中該雷射為一鎖模 雷射，且更包含一光調變器定位為自一鎖模雷射光脈衝序 列選擇該第一探測光脈衝。 2 5.如申請專利範圍第24項之裝置，其中該鎖模雷射光脈 衝序列和重該複測試圖為相位鎖定。 2 6.如申請專利範圍第1 9項之裝置，其中該第一參考光脈 衝源為一雷射。 2 7.如申請專利範圍第26項之裝置，其更包含一光調變器 且其中一輸出雷射束經該光調變器調變以提供該第一參考 光脈衝。 2 8.如申請專利範圍第1 9項之裝置，其中該等探測光脈衝 和該等參考光脈衝之波長大致相同。 2 9.如申請專利範圍第1 9項之裝置，其中在該測試圖之每 一重複中相對於該測試圖所提供第一參考光脈衝之時間為 固定。 3 0.如申請專利範圍第1 9項之裝置，其中該第一探測光脈 衝和該第一參考光脈衝沿一相同路徑導向至該半導體元件 3 1.如申請專利範圍第1 9項之裝置，其更包含一延遲路571108 6. Scope of Patent Application 2 2. The device according to item 19 of the scope of patent application, wherein the first detection light pulse source and the first reference light pulse source include the same laser. 2 3. The device according to item 19 of the patent application scope, wherein the first detection light pulse source is a laser. 24. The device of claim 23, wherein the laser is a mode-locked laser, and further comprises an optical modulator positioned to select the first detection light pulse from a mode-locked laser light pulse sequence. 2 5. The device according to item 24 of the scope of patent application, wherein the mode-locked laser light pulse sequence and the repeated test pattern are phase locked. 2 6. The device according to item 19 of the patent application scope, wherein the first reference light pulse source is a laser. 27. The device of claim 26, further comprising a light modulator, and an output laser beam is modulated by the light modulator to provide the first reference light pulse. 2 8. The device according to item 19 of the patent application scope, wherein the wavelengths of the detection light pulses and the reference light pulses are approximately the same. 2 9. The device according to item 19 of the scope of patent application, wherein the time relative to the first reference light pulse provided by the test chart in each repetition of the test chart is fixed. 30. The device according to item 19 of the patent application scope, wherein the first detection light pulse and the first reference light pulse are guided to the semiconductor element along a same path 3 1. The device according to item 19 of the patent application scope , Which also includes a delay path 第34頁 571108 六、申請專利範圍 > ^ °亥第一 ^木測光脈衝和該第二參考光脈衝沿該延遲路徑 向。 二 導向 32.如申請專利範圍第31項之裝置，其更包含一反饋迴路 :二延遲路徑耦接，藉以控制該延遲路徑之長度以補償半 導體元件之運動。 貝干 H申^專利範圍第32項之裝置’其中該反饋迴路使該 蒼考干涉信號穩定。 H申^專利範圍第19項之裝置，其中該偵測器包含一 使廿於=疋f u偵測該第一探測光脈衝和該第二探測光脈 乎拉啼 a k唬，一放大器耦接以接收該探測干 二‘ 2 β =叶器輕接以接收該放大的探測干涉信號，及 4 換器耦接以接收該累計的探測干涉信號。 ✓ 光偵制專利、乾圍第1 9項之裝置，其中該偵測器包含-衝並钤ΐ二位=偵測該第一參考光脈衝和該第二參考光脈 涉作^ 芩考干涉信號，一放大器耦接以接收該參考干 一1"μ汁态耦接以接收該放大的參考干涉信號，及 轉換為、耦接以接收該累計的參考干涉信號。 •如申請專利範圍第i 9 裝置一b 位以使一探測光脈衝之一 轉向，一匕3刀肌為疋 測該部分探測伞sr 4 ^ 先偵測裔定位以偵 技、，4 铋先脈衝並輸出一光偵測器信號，一放大哭耦 接以接收該光偵測哭严啼一罗斗1 放大杰輛 光偵測哭作鲈月：唬，一 π “、耦接以接收該放大的 測器信^虎及一數位轉換器搞接以接收該累計的光偵 37.如申請專利範圍第19項之裝置，其更包含一分流器定Page 34 571108 VI. Scope of patent application > The first photometric pulse and the second reference light pulse are directed along the delay path. Second guidance 32. The device according to item 31 of the patent application scope, further comprising a feedback loop: two delay paths are coupled to control the length of the delay path to compensate for the movement of the semiconductor element. The device of Bekan H's patent application No. 32, wherein the feedback loop stabilizes the Cangkao interference signal. The device of claim 19 in the patent scope, wherein the detector includes a detector for detecting the first detection light pulse and the second detection light pulse, and an amplifier is coupled to Receiving the detection stem '2 β = the leaf is lightly connected to receive the amplified detection interference signal, and the 4 switch is coupled to receive the accumulated detection interference signal. ✓ The device of patented light detection system, No. 19, where the detector includes-rush and two-bit = detection of the first reference light pulse and the second reference light pulse involve ^ test interference Signal, an amplifier is coupled to receive the reference stem 1 " μ juice state coupled to receive the amplified reference interference signal, and converted to, coupled to receive the accumulated reference interference signal. • If the i 9th device of the scope of the patent application applies a b-bit to turn one of the detection light pulses, a dagger 3 muscle is used to detect the part of the detection umbrella sr 4 ^ First detect the locator for detection, 4 bismuth first Pulse and output a light detector signal, a magnified couple is coupled to receive the light to detect the cry and cried a fight 1 to magnify the light detector to detect the crying month: bluff, a π ", coupled to receive the The enlarged sensor signal and a digital converter are connected to receive the accumulated light detection. 37. For the device in the 19th scope of the patent application, it further includes a shunt 第35頁 571108 六、申請專利範圍 位以使一參考光脈衝之一部分轉向，一光偵測器定位以偵 測該部分參考光脈衝並輸出一光偵測器信號，一放大器耦 接以接收該光偵測器信號，一累計器耦接以接收該放大的 光偵測器信號，及一數位轉換器耦接以接收該累計的光偵 測器信號。Page 35 571108 6. Apply for a patent to set a part of a reference light pulse to turn, a light detector is positioned to detect the part of the reference light pulse and output a light detector signal, an amplifier is coupled to receive the A photodetector signal, an accumulator is coupled to receive the amplified photodetector signal, and a digital converter is coupled to receive the accumulated photodetector signal. 第36頁Page 36